| With Computed Tomography(CT)widely applied to clinical diagnosis and treatment,the radiation dose of CT damage to human health more and more get the attention of the medical industry.How to reduce the patient's radiation dose and exert the diagnostic value of CT has become the research focus at home and abroad under the premise of ensuring the image quality.This paper systematically expounds the principles of physics and the thought of image reconstruction,and analyzes the traditional algorithms of FBP(filtered back projection).The problems of traditional algorithm such as high image noise and radiation dose are pointed out.In this paper,noise reduction algorithm based on iterative reconstruction is proposed.(1)According to the data stream of CT image reconstruction,an iterative algorithm based on the front projection domain and the hybrid projection domain is proposed from the front projection of the original data and the back projection of the image.(2)Based on the maximum likelihood method,an iterative algorithm based on the pre-projection domain is added with the wavelet transform factor to improve the high-frequency noise of the image and to reduce the partial differential of the anisotropic diffusion of the low-frequency noise.(3)Shrinking high-frequency part of the noise at the same time,but also improves the sharpness of the edge of the image.At the same time,the CT image reconstruction algorithm based on model iteration is studied.The CT system model,the geometric model of photon transmission and the iterative algorithm are combined to make use of the evaluation function and the new regularization term to achieve the goal of noise control.Clinical images demonstrate a significant improvement in the level of noise improvement.Through the use of iterative reconstruction algorithm,the image quality can be greatly optimized and the radiation dose of the patient reduced,which makes the clinical application of CT more abundant and can provide more clinical value for diagnosis. |
PDF Full Text Request